CAPÍTULO IV GEOMETRÍA PROYECTIVA:
OTROS MODELOS PARA LAS GEOMETRÍAS NO EUCLIDIANAS
This section and the next present the findings of our empirical research. In this section, we explain our analysis of the survey data and our hypotheses, and why we believe our approach did not bring about a change in immigration attitudes. We then present the main focus group findings
thematically. Finally, we draw some conclusions, focused particularly on how the nature of the debate on immigration might be improved through a greater understanding of the drivers of opinion and use of evidence.
3.1. Survey questions and experimental design
The data collected through the surveys and focus groups were analysed quantitatively and
qualitatively. Our quantitative analysis focused on changes in responses on the different questions between the three points at which participants filled out survey: at recruitment, immediately after having watched the video in the focus group, and then again two weeks after the focus group. In
what follows, we will refer to these time-points as pre, post, and follow-up, respectively. All
reliability estimates are provided at the initial time-point (pre).
Several of the items on the questionnaire were combined into scales. Specifically, three questions gauging participants’ agreement with statements to the effect that EU migrants take jobs from British workers, drive down wages for British workers, and have helped create jobs in the UK were
combined (after reverse scoring items where necessary) to construct a scale for economic
perceptions, with higher scorers being more positive about the economic impact of EU immigration (alpha = 0.76). Four questions gauging participants’ agreement with statements to the effect that EU migrants contribute more in taxes than they spend on public services, have a negative impact on the NHS, have a negative impact on schools, and have a negative impact on housing, were combined (again, after having reverse scored items where necessary) to construct a scale for fiscal perceptions, with higher scores being more positive about the fiscal impact of EU immigration on the UK (alpha = 0.86).
The primary form of analysis was non-parametric, given the ordinal nature of the items and scales. Specifically, overall longitudinal changes (ignoring experimental conditions) were analysed by way of Wilcoxon signed rank tests with continuity correction. The effects of individual experimental
conditions were analysed by calculating the difference in response between pre and post,
representing short-term effects, and separately for post and follow-up, representing longer-term
effects, and then running a Kruskal-Wallis rank sum test to identify any significant differences between any of the conditions, with a Dunn’s multiple comparison test for any post hoc analysis. As noted at the outset, our main aim was to investigate whether policy preferences could be moved if participants were informed about the economic impact of immigration following interventions controlling for well-known biases. This meant that, in addition to aforementioned scales of fiscal and economic perceptions, one item was of particular interest to us, namely: ‘The Government should use Brexit as an opportunity to cut down EU immigration’ (1 = Strongly disagree; 2 = Disagree; 3 = Neither agree nor disagree; 4 = Agree; 5 = Strongly agree). This was our main policy item, which also can be expected to tap into the sentiments that crop up in large-scale surveys regarding the desire on the part of people in the UK to reduce the number of immigrants. As such, we can formulate the following hypotheses:
Listening Hypothesis: Following an intervention designed to make people feel listened to and the subsequent provision of information about the economic and fiscal impact of EU
immigration on the UK, we should see significant movement (i) in a positive direction in
regards to economic and fiscal perceptions, and (ii) away from agreement with the statement
that the Government should use Brexit as an opportunity to cut down EU immigration, compared to the control group.
Biased Assimilation Hypothesis: Following an intervention designed to reduce biased
assimilation of evidence and the subsequent provision of information about the economic and
fiscal impact of EU immigration on the UK, we should see significant movement (i) in a positive
direction in regards to economic and fiscal perceptions, and (ii) away from agreement with the
statement that the Government should use Brexit as an opportunity to cut down EU immigration, compared to the control group.
In-group Favouritism Hypothesis: Following an intervention designed to reduce in-group bias
and the subsequent provision of information about the economic and fiscal impact of EU
immigration on the UK, we should see significant movement (i) in a positive direction in
regards to economic and fiscal perceptions, and (ii) away from agreement with the statement
that the Government should use Brexit as an opportunity to cut down EU immigration, compared to the control group.
3.2. Survey results
As we explained above, we surveyed focus group participants at three points in the research: pre,
post, and follow-up. In this section, we present the findings of the survey carried out at the pre
stage. The next section will then discuss any changes to the post and follow-up stage.
As indicated by our two policy survey items, participants wanted to see the Government use Brexit as an opportunity to cut down on EU immigration (median response of 4; 1 = Strongly Disagree, 2 = Agree, 3 = Neither Agree nor Disagree, 4 = Agree, 5 = Strongly Agree) and for the UK to rely on its own skill rather than on EU immigration (median response of 4).
Our scales for economic and financial perceptions suggested fairly neutral views on the economic impact of EU immigration (median score of 3 on a scale from 1 to 5, with higher numbers being more positive) and slightly more negative perceptions on fiscal impact (median score of 2.5).
There were some concerns about there being too many EU immigrants in the UK (median response of 4; 1 = Strongly Disagree, 2 = Agree, 3 = Neither Agree nor Disagree, 4 = Agree, 5 = Strongly Agree). Interestingly, participants seemed less concerned when asked whether there were too many EU immigrants in their local area (median response of 3).
The survey did not suggest that participants were particularly concerned about there being too many cultures coming into the country (median response of 3; 1 = Strongly Disagree, 2 = Agree, 3 = Neither Agree nor Disagree, 4 = Agree, 5 = Strongly Agree).
3.3. Movements in responses over time
We looked for any changes in responses over time that might indicate that the focus group methods and discussions had an impact on participants’ attitudes. A summary of changes in responses over
reporting responses at the pre-stage in terms of the median response. However, we have reported
changes in responses across time in terms of means11.
The first thing to note is that we see a statistically significant movement in a positive direction from
pre (i.e., at recruitment) to post (i.e., after the participants had seen the video) on our scales for
economic and fiscal perceptions, together with a significant negative movement, signifying a
movement away from agreement with the statement that the Government should use Brexit as an
opportunity to cut down on immigration. This is consistent with the participants having been moved in a positive direction as a result of the information provided in the video, and consequently
experiencing a shift in their policy preferences regarding numbers. This shift then waned in the period from post video to the follow-up survey two weeks later. However, it is not possible to ascribe these changes to participation in the focus groups since we lack a control group of people who did not attend.
The second thing to note is that we found no statistically significant differences in movements in attitudes between participants in the different focus group conditions, compared to the control group. Figure 6 offers an illustrative and representative example of the degree of similarity we saw across conditions with respect to any movements in responses on our main policy item, in this case
from pre to post.
Listening
Devil’s advocate
Proposal defence
Control
Less agreement No difference More agreement
Figure 6. Change in participants’ agreement with the statement ‘The Government should use Brexit as an opportunity to cut down on immigration’ from recruitment (pre) to after
having watched the video (post).
11 The reason for this is that the changes are small, and would tend to be obscured if reported in terms of
changes in medians. For that reason, means offer a better sense of the central tendency of movements across time, although their exact magnitude of course should be taken with a grain of salt, given the ordinal nature of the responses. Note, moreover, that none of the statistical tests have been performed on mean values—again, respecting the fact that our data is ordinal.
Given the absence of any statistically significant differences across conditions, we are unable to reject the null hypotheses corresponding to our three hypotheses above (Section 3.1.). It is possible that this is due to the relatively small samples and resulting lack of power. Alternatively, the
interventions did not make any difference. While our experiment doesn’t settle the matter, existing evidence suggests that the latter is the more likely possibility between the two. In other words, given that attempts prior to ours to move people’s policy preferences by providing information about its fiscal, and more generally economic, impact have failed to do so, one might reasonably look at our results as another instance of this happening.
At the same time, the focus groups generated rich and detailed qualitative data on the opinions, views and perspectives of a set of people on a major policy question at an historic point in time for the UK. In this respect, the focus group discussions go beyond much existing research by potentially offering explanations for why people hold particular views, rather than simply describing the views that they hold. The findings also suggest ways in which public understanding about immigration might be improved and how concerns and misconceptions might be addressed. We now turn to these findings.